[论文解读] UrbanGS: A Scalable and Efficient Architecture for Geometrically Accurate Large-Scene Reconstruction

Changbai Li, Haodong Zhu+7 more|arXiv (Cornell University)|Feb 2, 2026

Advanced Vision and Imaging被引用 0

一句话总结

UrbanGS 引入深度一致性 D-Normal 正则化、基于深度的自适应监督，以及空间自适应高斯修剪，以实现高保真、内存高效的城市尺度3D高斯散点重建，且具可扩展分区化。

ABSTRACT

While 3D Gaussian Splatting (3DGS) enables high-quality, real-time rendering for bounded scenes, its extension to large-scale urban environments gives rise to critical challenges in terms of geometric consistency, memory efficiency, and computational scalability. To address these issues, we present UrbanGS, a scalable reconstruction framework that effectively tackles these challenges for city-scale applications. First, we propose a Depth-Consistent D-Normal Regularization module. Unlike existing approaches that rely solely on monocular normal estimators, which can effectively update rotation parameters yet struggle to update position parameters, our method integrates D-Normal constraints with external depth supervision. This allows for comprehensive updates of all geometric parameters. By further incorporating an adaptive confidence weighting mechanism based on gradient consistency and inverse depth deviation, our approach significantly enhances multi-view depth alignment and geometric coherence, which effectively resolves the issue of geometric accuracy in complex large-scale scenes. To improve scalability, we introduce a Spatially Adaptive Gaussian Pruning (SAGP) strategy, which dynamically adjusts Gaussian density based on local geometric complexity and visibility to reduce redundancy. Additionally, a unified partitioning and view assignment scheme is designed to eliminate boundary artifacts and optimize computational load. Extensive experiments on multiple urban datasets demonstrate that UrbanGS achieves superior performance in rendering quality, geometric accuracy, and memory efficiency, providing a systematic solution for high-fidelity large-scale scene reconstruction.

研究动机与目标

• 解决城市环境中大尺度3D高斯散点法（3DGS）的几何不一致性与内存低效问题。
• 提出深度一致性 D-Normal 正则化框架，通过深度与法线线索更新所有高斯参数（位置与旋转）。
• 引入空间自适应高斯剪枝（SAGP）策略，基于局部几何与可见性降低冗余。
• 设计分区与视角分配方案，防止边界伪影并实现可扩展的多GPU重建。
• 在城市尺度数据集上展示最先进的渲染质量、几何精度与训练效率。

提出的方法

• 从深度图梯度派生 D-Normal 的深度一致性正则化，并以伪法线先验监督渲染法线。
• 深度一致性正则化，将单目深度锚点与稀疏SfM点对齐，使用反深度损失和几何感知置信权重。
• 一个统一损失 L_total，结合 RGB 重建损失、n-正则化、dn-正则化，以及以 w_d 加权的深度损失。
• 空间自适应高斯剪枝（SAGP），基于局部几何、可见性和视角频率为每个体素计算重要性，并剪除冗余高斯分布。
• 基于体素的分区与边界保持的视角分配，实现大尺度并行重建，降低边界伪影。
• 一种先于区块级训练的 SAGP 剪枝前置的分区策略，以实现城市尺度的可扩展性重建。

Figure 2: UrbanGS training pipeline and core components. (a) Training Pipeline: Starting from coarse global Gaussians, we apply spatially adaptive Gaussian pruning to obtain compact priors, contract and partition the scene into blocks, assign camera views using geometric and SSIM-based criteria, and

实验结果

研究问题

• RQ1 如何联合利用深度与法线线索，对大尺度场景中的3D高斯参数进行整体优化？
• RQ2 深度感知监督与自适应权重是否能提升城市尺度重建中的多视深度对齐与几何一致性？
• RQ3 如何高效剪枝高斯原语，以在内存效率和渲染/几何保真之间取得平衡？
• RQ4 哪种分区与视角分配策略能在多GPU设置下实现无边界、可扩展的大尺度重建？
• RQ5 UrbanGS 在渲染质量、几何精度与训练效率方面是否优于现有的大尺度3DGS方法？

主要发现

• UrbanGS 在多个城市数据集上实现了最先进的渲染与几何质量，超越若干大尺度基线方法。
• 深度一致性 D-Normal 正则化实现了对高斯位置与旋转的整体优化，提升表面保真度。
• 将深度一致性与反深度锚点结合、并加入几何感知置信项，提升跨视角深度对齐。
• 空间自适应高斯剪枝（SAGP）在降低内存与训练时间的同时，保留前景与远处区域的细节。
• 分区与边界保持的视角分配使城市尺度场景的并行训练更加高效。
• 消融研究显示 D-Normal 正则化与深度一致性组件带来显著增益，SAGP 与分区策略同样有效（如 PSNR/SSIM 增益及内存降低）。

Figure 3: Qualitative results of ours and other methods in image rendering on Mill-19 (Yu et al., 2022 ) and Urbanscene3D (Lin et al., 2022 ) .